Abstract: Under the conditions of fuzzy objective and fuzzy design constraints, the percentile formulation design rules for fuzzy robust design, which brings both the robustness of quality characteristic and the feasibility robustness of design constraints into account, are proposed when the product quality possesses nominal-the-best or larger-the-better or smaller-the- better characteristic. The optimal model of fuzzy robust designed with different types of quality characteristic is presented based on the proposed rules. A framework for solving the presented model, which combines the general optimal algorithm with the most probable point based percentile search algorithm, is also proposed. A case study is described to show the effectiveness of the proposed approach.

Key words: Fuzzy robust design, Larger-the-better characteristic, Nominal-the-best characteristic, Percentile, Smaller-the-better characteristic